ISS 2.1: The Computational Sensing Revolution in Array Processing
Recent advances in inverse problems, including sparse signal recovery and non-convex optimization have shifted the design paradigm for sensing systems. Computational methods have become an integral part of the design toolbox, enabling the use of algorithms to address some of the hardware challenges in designing such systems. One of the most promising applications of this paradigm shift has been in array imaging systems, such as ultrasonic, radar and optical (LIDAR). The impact is also timely, as array processing is becoming increasingly important in a variety of applications, including robotics, autonomous driving, medical imaging, and virtual reality, among others. This has led to continuous improvements in sensing hardware, but also to increasing demand for theory and methods to inform the system design and improve the processing. This talk will present a general inverse problem framework for array processing systems, which allows us to describe both the acquisition hardware and the scene being acquired. Under this framework we can exploit prior knowledge on the scene, the system, and the nature of a variety of errors that might occur, allowing for significant improvements in the reconstruction accuracy. Furthermore, we can consider the design of the system itself in the context of the inverse problem, leading to designs that are more efficient, more accurate, or less expensive, depending on the application. We will explore applications of this model to LIDAR and depth sensing, radar and distributed radar, and ultrasonic sensing. In the context of these applications, we will describe how different models can lead to improved specifications in radar and ultrasonic systems, robustness to position and timing errors in distributed array systems, and cost reduction and new capabilities in LIDAR systems.
Speaker: Dr. Petros T Boufounos, Senior Principal Research Scientist, Mitsubishi Electric Research Laboratories (MERL), USA
ISS 2.2: Application of Computer Vision Technologies to Insurance Business
In this talk, the speaker will describe the research and development work on computer vision, done at PingAn Technology, with applications to a few different scenarios related to insurance business. The talk will present the technical details of some advanced deep learning approaches and share with the audience the research direction and scope in the areas of image and video understanding related to insurance industry. The representative application examples presented in this talk include car damage estimation, pig identification, AI painting and sculpture, crop identification and lesion segmentation from medical images.
Speaker: Dr. Mei Han, PingAn Technology, US Research Lab at Silicon Valley
ISS 2.3: Applications of Image &Video Analysis in Media Industry
This talk will be devoted to emerging technologies of audiovisual analysis in media industry that pursue a progressive approach to video understanding constructed over a layer of hierarchical segmentation. The first objective is to achieve a reliable segmentation of video that can divide the sequence into scenes, shots, sentences, words, etc. This hierarchical segmentation may be viewed as a first step to video understanding but, by itself, it has a lot of possible applications that have not yet been exploited by the current media industry.The technologies involved in the analysis combine speech/non-speech detection, speech-to-text conversion, face detection and recognition, text detection, object and places identification, scene detection, etc. The performance of these technologies with real video will be briefly reviewed in the talk. Also, some benefits for the media industry, increasing the quality of services, will be explored.
As a case study, this talk presents the related techniques developed for media companies at Ugiat Technologies, a spin-off from the UPC Barcelona Tech. by including smArDS, a software tool that analyzes video and proposes the best locations for inserting ads. The system also provides context metadata for selecting the most appropriate ad in each location (i.e. If the system detects people taking coffee in a video scene then it will propose a Bar&Restaurants ad).
The talk will also present other commercial applications using similar technologies like intelligent players, automatic video summarization, advertisement detection and replacement, detection of clips in news programs, etc.
Speaker: Dr. Francesc Tarrés, CEO, Ugiat & UPC